Abstract: In order to popularize the diagnostic knowledge of cotton pests and meet the cotton farmers’ need for domain experts, this paper presents a web-based pests diagnosis system for cotton by using the technology of Java, Internet and the database development tools of SQL Server 2000. The research achieved the goal of remote diagnosis of cotton pests with the help of internet. With the methods of fuzzy inference and case gaining, the system improved traditional knowledge acquisition methods, increased the diagnosis accuracy, and offset the low efficiency of the diagnosis with single method of case retrieval. The system consists of the functions of knowledge browsing, case retrieval and expert consultation, which can provide not only fast, convenient and accurate intelligence enquiry service, but also remote online diagnosis in virtue of expert consultation for cotton farmers.